/Data-Science-Competition

Random Forest and SVM for Multiclass Classification.

Primary LanguageJupyter Notebook

Data-Science-Competition

In October 2021, I participated in the DataScience Competition with IE.

We were given Dataset of space objects with several csv's to merge and were completely free of chosing the algorithms we wanted to go with. We were to classify every objects into either "Quasar", "Star", or "Galaxy".

I've coded from scratch a simple yet efficient Random Forest which gave me a ~91% accuracy, and then created a Support Vector Machines algorithms with Dataiku which gave me a ~94% accuracy.

I ended top 30% of the competition.